Hepatocellular carcinoma is the one of most frequent causes of death by cancer in the world. There is no reliable diagnosis prior to late stages of disease and no hope for cure except surgery. The overall goal of the proposal is to discover and identify distinctive alterations of protein expression in early precancerous lesions isolated from their physiological microenvironment. The data obtained will provide early molecular markers with diagnostic and therapeutic potential for early intervention in human liver cancer. We propose to apply innovative proteomics and laser capture microdissection microscopy (LCM) to study a well-characterized animal model of liver carcinogenesis (RH) that exhibits well-defined, synchronous stages of initiation and progression of liver cancer that are strikingly similar to those in liver cancer in humans. In the R21 phase, we will focus on demonstrating our ability to generate useful patterns/profiles with combined LCM and protein technologies using control tissue/serum and one preneoplastic stage of the biological model system. In the R33 phase, we will apply the pattern/profile generation technologies to the cells and sera of each of the early stages and control paradigms so that we can identify candidate markers and targets. We will use both global and targeted proteomics strategies to: 1) identify difference proteins in cells early in the development of liver cancer; 2) identify unique serum markers at early stages; 3) determine progressive changes in tubulin isotype composition, which is associated with development of drug resistance; and 4) identify changes in proteins associated with the cytoskeletal scaffold.